Yes, AI Is Really Impacting The Job Market. Here’s What To Do.Josh Bersin 在 2025 年末指出,美国就业市场正在出现结构性变化。整体失业率上升至 4.6%,其中应届大学毕业生的失业率接近 10%,成为最受冲击的群体。与此同时,不要求大学学历的岗位持续增长,一线员工的重要性正在被重新定义。更值得关注的是信任问题。Edelman 调研显示,70% 的员工不信任企业关于 AI 裁员的说法,只有 27% 信任 CEO。AI 不只是技术工具,而是一场社会与组织层面的转型。Josh Bersin 强调,AI 并非消灭岗位,而是放大能力。真正的挑战在于,企业是否愿意投资年轻人才,是否能用透明沟通化解 AI 焦虑。
详细来看
All year I’ve been studying the employment data and talking with press about the smallish impact of AI on the job market. Most of the slowdown in US jobs, from my data and conversations, has been driven by cost-cutting and general economic uncertainty, not explicit AI job replacement.
Well going into 2026 the situation is changing. The US unemployment rate is now 4.6%, up from 4.2% one year ago (a 9.5% increase) and 3.7% in November of 2023 (a 24.7% increase in two years). These are significant increases, especially considering that unemployment was 3.6% in November of 2022.
This tells me that the US economy is slowing after the post-pandemic “revenge buying” frenzy of 2021 and 2022. And of course US tariffs, inflation, and relatively high interest rates all contribute.
But now let’s look under the covers and break out unemployment into two sub-groups: new college graduates (24 years and younger), and more seasoned workers (age 25-35). Suddenly you see a divergence.
The green line, tenured college graduates, shows a steady unemployment rate below the average. This makes sense: these are experienced employees with skills, judgement, and seasoned decision-making maturity.
The orange line, new college graduates, is trending upward. In fact right now it’s almost 10%, which is the highest it has been since July 2021, the peak recovery from the pandemic. Looking backwards, the only time young college grad unemployment was this high was in 2011, a period of recovery from the 2008 recession. (St. Louis Fed agrees.)
And by the way, to round this out, jobs that do not require a degree are plentiful, roughly 82% of the workforce (up from 79% five years ago). So AI is not only slowing new college grad hiring, it’s also reducing the total number of jobs that require college.
There are three important things happening here:
First, whether it’s correct or not, employers are slowing down entry level hiring. Companies hire new college grads for many reasons (largely for talent pipeline), and many newly minted grads are far more AI-ready than we are. Despite this, it appears to economists that it’s harder than ever for these young folks to compete, so they need to “sell” their AI readiness and learning capacity.
Second, the frontline workforce is becoming much more important. The general automation of white collar work (it’s still early days) and the explosion of jobs in healthcare, social services, retail, repair, entertainment and distribution are making the “college grad” part of the workforce relatively smaller. That’s not to say the money isn’t good, but as a CEO or leader more and more of your energy has to go into supporting these frontline workers. (Read our Frontline-First research for more.)
Third, employees don’t trust CEO talk about jobs. A new study by Edelman shows a massive lack of people’s trust in business leaders (and AI scientists) around AI. This 5,000+ worker survey found that 70% of US workers do not trust statements about AI job reductions. When asked “who you do trust” only 27% of US workers trust the CEO. So we, as leaders, have a trust problem.
Here’s the trust data, and this is all about “Trust in AI’s Value” not “trust in the AI platform.”
AI Is a Socio-Technological Innovation
As I talked about in this week’s podcast, AI is “socio-technological.” It has many societal and sociological impacts.
If only half your employees believe what leaders are telling them, they’re going to hold back, grumble, and resist change. This is why economic insecurity is high: people are concerned about their jobs, careers, and future earnings.
(So AI anxiety could actually lower economic productivity!)
The solution to this is not to ignore the topic, but rather to discuss it openly.
None of us really know how much impact AI will have (I do know most platforms over-sell its value right now), and AI is a little scary.
We have to get comfortable with phones that talk back to us, creepy emails that know our name, avatar-based job interviews, AI-driven career advice, and AI-informed performance reviews. And in 2026 we’re going to see digital twins, robots, and more real-life animations of people at work. (Galileo Learn uses a “Josh Agent” to coach and challenge you as you learn.)
Here’s my advice. If you’re holding back on entry-level hiring you may be making a mistake. Younger staff, who have lived with this technology for more of their lives, are likely to be the ones to most quickly use AI, build with AI, and innovate with its new applications.
People who are tenured tend to see new tools as a way to “speed up what they know how to do.” New employees might just say “why not do it this way?” and bring you the reinvention you need.
Everyone Has The Opportunity To Be A Superworker Now
AI is not a job killer, it’s a big job-leveler.
You, as a younger worker, have access to information and research which was often hoarded by experts. If you’re willing to roll up your sleeves, you can move from “apprentice” to “newly minted expert” quickly. And if you’re looking for a job there’s no excuse for not becoming an expert on the company before you talk with a recruiter.
For senior, more tenured people the same applies. You can’t rely on your experience alone any more: you, too, should be digging in and learning about new technologies, tools, and advancements in your domain.
Employers: Be Careful How You Think
For hiring managers and executives, beware of the “tenure trap” above. Just because a senior person knows your business better, you may find that the young “AI-Guru” right out of college catches up fast. Remember, tenured people may see AI as a way to “do things the old way faster” rather than “rethink the way we work.”
For HR leaders and recruiters, remember one thing. Younger workers may learn faster and ultimately improve productivity at a faster rate (plus they cost less). If you seek out fast-learning AI pioneers they could be your Superworkers of the future.
And for CEOs and other execs, be honest and thoughtful about your plans. All our research points to AI as a “scaling technology,” not one to “eliminate jobs.” The more honest and supportive you are, the faster your employees will adapt and help your company stay ahead.
20州联合起诉10万美元 H-1B 签证费:一场牵动美国教育与医疗体系的人才之争美国 20 个州正式起诉特朗普政府,反对 10 万美元 H-1B 新签证费用。州政府指出,该政策未经国会授权,将严重冲击教育、医疗等公共领域的人才供给。案件已进入联邦法院,可能对未来美国用工与移民政策产生深远影响,值得 HR 和企业持续关注。
美国围绕 H-1B 签证的政策争议再次升级。近日,包括加利福尼亚州和马萨诸塞州在内的 20 个州,正式向联邦法院提起诉讼,挑战特朗普政府对新 H-1B 签证征收 10 万美元费用 的决定。这一罕见的多州联合行动,迅速在政界、学界和雇主群体中引发强烈反响。
本次诉讼由加州总检察长 Rob Bonta 和马萨诸塞州总检察长 Andrea Joy Campbell 牵头。起诉方核心观点在于:该费用并非由国会授权设立,而是行政部门单方面决定,违反了联邦法律,也背离了国会当初设立 H-1B 项目的立法初衷。
州政府认为,H-1B 项目的本质是帮助美国在关键专业领域补充本土难以满足的人才缺口,而并非设置“准入门槛”式的高额收费。10 万美元的签证费用,不仅远超实际行政处理成本,也将对公共部门造成实质性伤害。
从数据来看,这一担忧并非空穴来风。教育行业是 H-1B 签证使用量排名第三的职业领域,全美约有 3 万名教育工作者 依赖该签证体系。在医疗领域,仅 2024 财年,就有 近 1.7 万名 医疗与健康相关从业者获得 H-1B 签证,其中约一半为医生和外科医生。根据官方预测,如果缺乏海外医疗人才补充,美国到 2036 年可能面临 8.6 万名医生 的结构性短缺。
相比之下,目前雇主为一名 H-1B 员工承担的总费用,通常在 960 美元至 75,595 美元 之间,已经包含多项法定与合规成本。州政府指出,将费用一次性抬升至 10 万美元,不仅缺乏成本依据,更可能迫使公立大学、医院和学区直接放弃国际招聘。
对此,美国国土安全部(DHS)态度强硬。DHS 在回应中表示,该收费措施是合法的,并将本次诉讼描述为“民主党州总检察长出于政治动机,对总统移民政策的阻挠行为”。双方立场分化明显,也使案件的政治与制度意义进一步放大。
值得注意的是,反对该费用的不仅是州政府。美国商会(US Chamber of Commerce) 以及 美国大学协会(Association of American Universities) 也已就同一问题另行提起诉讼,显示出企业界和高等教育体系对这一政策的普遍担忧。
在更广泛的背景下,特朗普政府此前还宣布,将加强对 H-1B 申请人的审查,包括对社交媒体的系统性审阅。这意味着,H-1B 政策正在从“成本层面”和“审查层面”同时收紧,其外溢影响已不再局限于科技行业,而是扩散至教育、医疗等公共服务领域。
目前,该案件已以 State of California v. Noem 为名,提交至 马萨诸塞州联邦地区法院。无论最终判决结果如何,这场诉讼都将对美国未来高技能移民政策、公共部门用工能力以及全球人才流动格局产生深远影响。
对 HR 从业者、雇主和国际人才而言,这不仅是一项费用调整的争议,更是一场关于“美国是否仍愿意为关键行业引入全球人才”的制度性拷问。